DOCSLIB.ORG

GCN-Sem at Semeval-2019 Task 1: Semantic Parsing Using Graph Convolutional and Recurrent Neural Networks

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
GCN-Sem at Semeval-2019 Task 1: Semantic Parsing Using Graph Convolutional and Recurrent Neural Networks GCN-Sem at SemEval-2019 Task 1: Semantic Parsing using Graph Convolutional and Recurrent Neural Networks Shiva Taslimipoor Omid Rohanian Sara Mozeˇ Research Group in Computational Linguistics University of Wolverhampton, UK fshiva.taslimi, omid.rohanian, [email protected] Abstract automatically converted to other parsing formats, e.g. Abstract Meaning Representation (AMR) and This paper describes the system submitted to Semantic Dependency Parsing (SDP), inter alia the SemEval 2019 shared task 1 ‘Cross-lingual Semantic Parsing with UCCA’. We rely on the (Hershcovich et al., 2018). semantic dependency parse trees provided in Although the schemes are formally different, the shared task which are converted from the they have shared semantic content. In order to original UCCA files and model the task as tag- perform our experiments, we target the converted ging. The aim is to predict the graph structure CONLLU format, which corresponds to tradi- of the output along with the types of relations tional bi-lexical dependencies and rely on the con- among the nodes. Our proposed neural archi- tecture is composed of Graph Convolution and version methodology which is provided in the BiLSTM components. The layers of the sys- shared task (Hershcovich et al., 2019) to attain tem share their weights while predicting de- UCCA graphs. pendency links and semantic labels. The sys- UCCA graphs contain both explicit and implicit tem is applied to the CONLLU format of the units 1 However, in bi-lexical dependencies, nodes input data and is best suited for semantic de- are text tokens and semantic relations are direct pendency parsing. bi-lexical relations between the tokens. The con- 1 Introduction version between the two format results in partial loss of information. Nonetheless, we believe that Universal Conceptual Cognitive Annotation it is worth trying to model the task using one of (UCCA) (Abend and Rappoport, 2013) is a the available formats (i.e. semantic dependency semantically motivated approach to grammatical parsing) which is very popular among NLP re- representation inspired by typological theories of searchers. grammar (Dixon, 2012) and Cognitive Linguistics Typically, transition-based methods are used in literature (Croft and Cruse, 2004). In parsing, syntactic (Chen and Manning, 2014) and seman- bi-lexical dependencies that are based on binary tic (Hershcovich et al., 2017) dependency parsing. head-argument relations between lexical units By contrast, our proposed system shares several are commonly employed in the representation of similarities with sequence-to-sequence neural ar- syntax (Nivre et al., 2007; Chen and Manning, chitectures, as it does not specifically deal with 2014) and semantics (Hajicˇ et al., 2012; Oepen parsing transitions. Our model uses word, POS et al., 2014; Dozat and Manning, 2018). and syntactic dependency tree representations as UCCA differs significantly from traditional de- input and directly produces an edge-labeled graph pendency approaches in that it attempts to ab- representation for each sentence (i.e. edges and stract away traditional syntactic structures and re- their labels as two separate outputs). This multi- lations in favour of employing purely semantic label neural architecture, which consists of a BiL- distinctions to analyse sentence structure. The STM and a Graph Convolutional Network (GCN), shared task, ‘cross-lingual semantic parsing with is described in Section3. UCCA’ (Hershcovich et al., 2019) consists in pars- ing English, German, and French datasets using 1Explicit units (terminal nodes) correspond to tokens in the UCCA semantic tagset. In order to enable the text, but implicit (semantic) units have no corresponding multi-task learning, the UCCA-annotated data is component in the text. 102 Proceedings of the 13th International Workshop on Semantic Evaluation (SemEval-2019), pages 102–106 Minneapolis, Minnesota, USA, June 6–7, 2019. ©2019 Association for Computational Linguistics 2 Related Work of words). A recent trend in parsing research is sequence-to- Graph Convolution. Convolutional Neural Net- sequence learning (Vinyals et al., 2015b; Kitaev works (CNNs), as originally conceived, are se- and Klein, 2018), which is inspired from Neural quential in nature, acting as detectors of N- Machine Translation. These methods ignore ex- grams (Kim, 2014), and are often used as feature- plicit structural information in favour of relying on generating front-ends in deep neural networks. long-term memory, attention mechanism (content- Graph Convolutional Network (GCN) has been in- based or position-based) (Kitaev and Klein, 2018) troduced as a way to integrate rich structural rela- or pointer networks (Vinyals et al., 2015a). By do- tions such as syntactic graphs into the convolution ing so, high-order features are implicitly captured, process. which results in competitive parsing performance In the context of a syntax tree, a GCN can be (Jia and Liang, 2016). understood as a non-linear activation function f Sequence-to-sequence learning has been partic- and a filter W with a bias term b: X ularly effective in Semantic Role Labeling (SRL) c = f( W xi + b) (1) (Zhou and Xu, 2015). By augmenting these i2r(v) models with syntactic information, researchers where r(v) denotes all the words in relation with have been able to develop state-of-the-art systems a given word v in a sentence, and c represents the for SRL (Marcheggiani and Titov, 2017; Strubell output of the convolution. Using adjacency matri- et al., 2018). ces, we define graph relations as mask filters for As information derived from dependency parse the inputs (Kipf and Welling, 2017; Schlichtkrull trees can significantly contribute towards under- et al., 2017). standing the semantics of a sentence, Graph Con- In the present task, information from each graph volutional Network (GCN) (Kipf and Welling, corresponds to a sentence-level dependency parse 2017) is used to help our system perform semantic tree. Given the filter W and bias b , we can there- parsing while attending to structural syntactic in- s s fore define the sentence-level GCN as follows: formation. The architecture is similar to the GCN component employed in Rohanian et al.(2019) for C = f(W XT A + b ) (2) detecting gappy multiword expressions. s s s where Xn×v, An×n, and Co×n are tensor rep- 3 Methodology resentation of words, the adjacency matrix, and the convolution output respectively.3 In Kipf For this task, we employ a neural architecture util- and Welling(2017), a separate adjacency matrix ising structural features to predict semantic pars- is constructed for each relation to avoid over- ing tags for each sentence. The system maps a parametrising the model; by contrast, our model sentence from the source language to a probability is limited to the following three types of relations: distribution over the tags for all the words in the 1) the head to the dependents, 2) the dependents sentence. Our architecture consists of a GCN layer to the head, and 3) each word to itself (self-loops) (Kipf and Welling, 2017), a bidirectional LSTM, similar to Marcheggiani and Titov(2017). The fi- and a final dense layer on top. nal output is the maximum of the weights from the The inputs to our system are sequences of three individual adjacency matrices. words, alongside their corresponding POS and The model architecture is depicted in Figure1. named-entity tags.2 Word tokens are repre- sented by contextualised ELMo embeddings (Pe- 4 Experiments ters et al., 2018), and POS and named-entity tags Our system participated in the closed track for En- are one-hot encoded. We also use sentence-level glish and German and the open track for French. syntactic dependency parse information as input We exclusively used the data provided in the to the system. In the GCN layer, the convolu- shared task. The system is trained on the train- tion filters operate based on the structure of the ing data only, and the parameters are optimised us- dependency tree (rather than the sequential order ing the development set. The results are reported 2spaCy (Honnibal and Johnson, 2015) is used to generate 3o: output dimension; v: word vectors dimension; n: sen- POS, named-entity and syntactic dependency tags. tence length 103 Word as explained in Section 4.1. The parameters are Representation optimised on the English Wiki development data (batch-size = 16 and number of epochs = 100) and GCN GCN used for all four settings. As no training data was (Root to (Dependency GCN Dependency) to Root) (Root) available for French, the trained system on English Wiki was used to parse French sentences of 20K Max Leagues. For this reason the French model is eval- uated within the open track. BiLSTM 4.3 Official Evaluation Our model predicts two outputs for each dataset: FFN primary edges and their labels (UCCA semantic categories). 5 Figure 1: A GCN-based recurrent architecture. Table1 shows the performance (in terms of pre- on blind-test data in both in-domain and out-of- cision, recall, and F1-score) for predicting primary domain settings. We focus on predicting the pri- edges in both labeled (i.e. with semantic tags) and mary edges of UCCA semantic relations and their unlabeled settings (i.e. ignoring semantic tags). labels. Table2 shows F1-scores for each semantic cate- gory separately. Although the overall performance 4.1 Data of the system, as shown in the official evaluation in Table1, is not particularly impressive, there are The datasets of the shared task are devised for four a few results worth reporting. These are listed in settings: 1) English in-domain, using the Wiki cor- Table2. pus; 2) English out-of-domain, using the Wiki cor- pus as training and development data, and 20K Our system is ranked second in predicting four Leagues as test data; 3) German in-domain, using relations, i.e. L (linker), N (Connector), R (Rela- the 20K Leagues corpus; 4) French setting with tor), and G (Ground), in all settings displayed in no training data (except trial data), using the 20K bold.
Load more

Recommended publications
  • Semantic Role Labeling: an Introduction to the Special Issue
    Semantic Role Labeling: An Introduction to the Special Issue Llu´ıs Marquez` ∗ Universitat Politecnica` de Catalunya Xavier Carreras∗∗ Massachusetts Institute of Technology Kenneth C. Litkowski† CL Research Suzanne Stevenson‡ University of Toronto Semantic role labeling, the computational identification and labeling of arguments in text, has become a leading task in computational linguistics today. Although the issues for this task have been studied for decades, the availability of large resources and the development of statistical machine learning methods have heightened the amount of effort in this field. This special issue presents selected and representative work in the field. This overview describes linguistic background of the problem, the movement from linguistic theories to computational practice, the major resources that are being used, an overview of steps taken in computational systems, and a description of the key issues and results in semantic role labeling (as revealed in several international evaluations). We assess weaknesses in semantic role labeling and identify important challenges facing the field. Overall, the opportunities and the potential for useful further research in semantic role labeling are considerable. 1. Introduction The sentence-level semantic analysis of text is concerned with the characterization of events, such as determining “who” did “what” to “whom,” “where,” “when,” and “how.” The predicate of a clause (typically a verb) establishes “what” took place, and other sentence constituents express the participants in the event (such as “who” and “where”), as well as further event properties (such as “when” and “how”). The primary task of semantic role labeling (SRL) is to indicate exactly what semantic relations hold among a predicate and its associated participants and properties, with these relations Departament de Llenguatges i Sistemes Informatics,` Universitat Politecnica` de Catalunya, Jordi Girona ∗ Salgado 1–3, 08034 Barcelona, Spain.
    [Show full text]
  • Ten Years of Babelnet: a Survey
    Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence (IJCAI-21) Survey Track Ten Years of BabelNet: A Survey Roberto Navigli1 , Michele Bevilacqua1 , Simone Conia1 , Dario Montagnini2 and Francesco Cecconi2 1Sapienza NLP Group, Sapienza University of Rome, Italy 2Babelscape, Italy froberto.navigli, michele.bevilacqua, [email protected] fmontagnini, [email protected] Abstract to integrate symbolic knowledge into neural architectures [d’Avila Garcez and Lamb, 2020]. The rationale is that the The intelligent manipulation of symbolic knowl- use of, and linkage to, symbolic knowledge can not only en- edge has been a long-sought goal of AI. How- able interpretable, explainable and accountable AI systems, ever, when it comes to Natural Language Process- but it can also increase the degree of generalization to rare ing (NLP), symbols have to be mapped to words patterns (e.g., infrequent meanings) and promote better use and phrases, which are not only ambiguous but also of information which is not explicit in the text. language-specific: multilinguality is indeed a de- Symbolic knowledge requires that the link between form sirable property for NLP systems, and one which and meaning be made explicit, connecting strings to repre- enables the generalization of tasks where multiple sentations of concepts, entities and thoughts. Historical re- languages need to be dealt with, without translat- sources such as WordNet [Miller, 1995] are important en- ing text. In this paper we survey BabelNet, a pop- deavors which systematize symbolic knowledge about the ular wide-coverage lexical-semantic knowledge re- words of a language, i.e., lexicographic knowledge, not only source obtained by merging heterogeneous sources in a machine-readable format, but also in structured form, into a unified semantic network that helps to scale thanks to the organization of concepts into a semantic net- tasks and applications to hundreds of languages.
    [Show full text]
  • ACNLP at Semeval-2020 Task 6: a Supervised Approach for Definition
    ACNLP at SemEval-2020 Task 6: A Supervised Approach for Definition Extraction Fabien Caspani Pirashanth Ratnamogan Mathis Linger Mhamed Hajaiej BNP Paribas, France ffabien.caspani, pirashanth.ratnamogan, mathis.linger, [email protected] Abstract We describe our contribution to two of the subtasks of SemEval 2020 Task 6, DeftEval: Extracting term-definition pairs in free text. The system for Subtask 1: Sentence Classification is based on a transformer architecture where we use transfer learning to fine-tune a pretrained model on the downstream task, and the one for Subtask 3: Relation Classification uses a Random Forest classifier with handcrafted dedicated features. Our systems respectively achieve 0.830 and 0.994 F1-scores on the official test set, and we believe that the insights derived from our study are potentially relevant to help advance the research on definition extraction. 1 Introduction SemEval 2020 Task 6 (Spala et al., 2020) is a definition extraction problem divided into three different subtasks, in which the objective is to find the term-definition pairs in free text, possibly spanning across sentences boundaries. In Subtask 1 the system classifies a sentence as containing a definition or not. Subtask 2 consists of labeling each token with their corresponding BIO tags. Finally, in Subtask 3 the system pairs tagged entities, and labels the relation existing between them. The Definition Extraction from Texts (DEFT) corpus (Spala et al., 2019) is the dataset used for this task, and consists of human-annotated data from textbooks on a variety of subjects. Definition extraction methods in the research literature fall into one of the following categories.
    [Show full text]
  • Semeval 2020 Task 1: Unsupervised Lexical Semantic Change Detection
    SemEval 2020 Task 1: Unsupervised Lexical Semantic Change Detection November 11, 2020 Dominik Schlechtweg,| Barbara McGillivray,};~ Simon Hengchen,♠ Haim Dubossarsky,~ Nina Tahmasebi♠ [email protected] |University of Stuttgart, }The Alan Turing Institute, ~University of Cambridge ♠University of Gothenburg 1 Introduction I evaluation is currently our most pressing problem in LSC detection 1 I SemEval is competition-style semantic evaluation series I SemEval 2020 Task 1 on Unsupervised Lexical Semantic Change Detection (Schlechtweg, McGillivray, Hengchen, Dubossarsky, & Tahmasebi, 2020)2 I datasets for 4 languages with 100,000 human judgments I 2 subtasks I 33 teams submitted 186 systems 1https://semeval.github.io/ 2 https://languagechange.org/semeval/ 2 Tasks I comparison of two time periods t1 and t2 (i) reduces the number of time periods for which data has to be annotated (ii) reduces the task complexity I two tasks: I Subtask 1 { Binary classification: for a set of target words, decide which words lost or gained senses between t1 and t2, and which ones did not. I Subtask 2 { Ranking: rank a set of target words according to their degree of LSC between t1 and t2. I defined on word sense frequency distributions 3 Sense Frequency Distributions (SFDs) Figure 1: An example of a sense frequency distribution for the word cell in two time periods. 4 Corpora t1 t2 English CCOHA 1810-1860 CCOHA 1960-2010 German DTA 1800-1899 BZ+ND 1946-1990 Latin LatinISE -200{0 LatinISE 0{2000 Swedish Kubhist 1790-1830 Kubhist 1895-1903 Table 1: Time-defined subcorpora for each language. 5 Annotation 1.
    [Show full text]
  • Semeval-2013 Task 4: Free Paraphrases of Noun Compounds
    SemEval-2013 Task 4: Free Paraphrases of Noun Compounds Iris Hendrickx Zornitsa Kozareva Radboud University Nijmegen & University of Southern California Universidade de Lisboa [email protected] [email protected] Preslav Nakov Diarmuid OS´ eaghdha´ QCRI, Qatar Foundation University of Cambridge [email protected] [email protected] Stan Szpakowicz Tony Veale University of Ottawa & University College Dublin Polish Academy of Sciences [email protected] [email protected] Abstract The frequency spectrum of compound types fol- lows a Zipfian distribution (OS´ eaghdha,´ 2008), so In this paper, we describe SemEval-2013 Task many NC tokens belong to a “long tail” of low- 4: the definition, the data, the evaluation and frequency types. More than half of the two-noun the results. The task is to capture some of the types in the BNC occur exactly once (Kim and Bald- meaning of English noun compounds via para- phrasing. Given a two-word noun compound, win, 2006). Their high frequency and high produc- the participating system is asked to produce tivity make robust NC interpretation an important an explicitly ranked list of its free-form para- goal for broad-coverage semantic processing of En- phrases. The list is automatically compared glish texts. Systems which ignore NCs may give up and evaluated against a similarly ranked list on salient information about the semantic relation- of paraphrases proposed by human annota- ships implicit in a text. Compositional interpretation tors, recruited and managed through Ama- is also the only way to achieve broad NC coverage, zon’s Mechanical Turk. The comparison of because it is not feasible to list in a lexicon all com- raw paraphrases is sensitive to syntactic and morphological variation.
    [Show full text]
  • Detecting Semantic Difference: a New Model Based on Knowledge and Collocational Association
    DETECTING SEMANTIC DIFFERENCE: A NEW MODEL BASED ON KNOWLEDGE AND COLLOCATIONAL ASSOCIATION Shiva Taslimipoor1, Gloria Corpas2, and Omid Rohanian1 1Research Group in Computational Linguistics, University of Wolverhampton, UK 2Research Group in Computational Linguistics, University of Wolverhampton, UK and University of Malaga, Spain {shiva.taslimi, omid.rohanian}@wlv.ac.uk [email protected] Abstract Semantic discrimination among concepts is a daily exercise for humans when using natural languages. For example, given the words, airplane and car, the word flying can easily be thought and used as an attribute to differentiate them. In this study, we propose a novel automatic approach to detect whether an attribute word represents the difference between two given words. We exploit a combination of knowledge-based and co- occurrence features (collocations) to capture the semantic difference between two words in relation to an attribute. The features are scores that are defined for each pair of words and an attribute, based on association measures, n-gram counts, word similarity, and Concept-Net relations. Based on these features we designed a system that run several experiments on a SemEval-2018 dataset. The experimental results indicate that the proposed model performs better, or at least comparable with, other systems evaluated on the same data for this task. Keywords: semantic difference · collocation · association measures · n-gram counts · word2vec · Concept-Net relations · semantic modelling. 1. INTRODUCTION Semantic modelling in natural language processing requires attending to both semantic similarity and difference. While similarity is well-researched in the community (Mihalcea & Hassan, 2017), the ability of systems in discriminating between words is an under-explored area (Krebs et al., 2018).
    [Show full text]
  • Semeval-2017 Task 2: Multilingual and Cross-Lingual Semantic Word Similarity
    SemEval-2017 Task 2: Multilingual and Cross-lingual Semantic Word Similarity Jose Camacho-Collados*1, Mohammad Taher Pilehvar*2, Nigel Collier2 and Roberto Navigli1 1Department of Computer Science, Sapienza University of Rome 2Department of Theoretical and Applied Linguistics, University of Cambridge 1 collados,navigli @di.uniroma1.it { 2 mp792,nhc30}@cam.ac.uk { } Abstract word representation, a research field that has re- cently received massive research attention mainly This paper introduces a new task on Multi- as a result of the advancements in the use of neural lingual and Cross-lingual Semantic Word networks for learning dense low-dimensional se- Similarity which measures the semantic mantic representations, often referred to as word similarity of word pairs within and across embeddings (Mikolov et al., 2013; Pennington five languages: English, Farsi, German, et al., 2014). Almost any application in NLP that Italian and Spanish. High quality datasets deals with semantics can benefit from efficient se- were manually curated for the five lan- mantic representation of words (Turney and Pan- guages with high inter-annotator agree- tel, 2010). ments (consistently in the 0.9 ballpark). However, research in semantic representation These were used for semi-automatic con- has in the main focused on the English language struction of ten cross-lingual datasets. 17 only. This is partly due to the limited availabil- teams participated in the task, submitting ity of word similarity benchmarks in languages 24 systems in subtask 1 and 14 systems in other than English. Given the central role of subtask 2. Results show that systems that similarity datasets in lexical semantics, and given combine statistical knowledge from text the importance of moving beyond the barriers of corpora, in the form of word embeddings, the English language and developing language- and external knowledge from lexical re- independent and multilingual techniques, we felt sources are best performers in both sub- that this was an appropriate time to conduct a task tasks.
    [Show full text]
  • Semeval-2016 Task 1: Semantic Textual Similarity, Monolingual And
    SemEval-2016 Task 1: Semantic Textual Similarity, Monolingual and Cross-Lingual Evaluation Eneko Agirrea, Carmen Baneab, Daniel Cerd, Mona Diabe, Aitor Gonzalez-Agirrea, Rada Mihalceab, German Rigaua, Janyce Wiebef aUniversity of the Basque Country bUniversity of Michigan Donostia, Basque Country Ann Arbor, MI dGoogle Inc. eGeorge Washington University f University of Pittsburgh Mountain View, CA Washington, DC Pittsburgh, PA Abstract ranges from complete semantic equivalence to com- plete semantic dissimilarity. The intermediate levels Semantic Textual Similarity (STS) seeks to capture specifically defined degrees of partial sim- measure the degree of semantic equivalence ilarity, such as topicality or rough equivalence, but between two snippets of text. Similarity is ex- with differing details. The snippets being scored are pressed on an ordinal scale that spans from approximately one sentence in length, with their as- semantic equivalence to complete unrelated- ness. Intermediate values capture specifically sessment being performed outside of any contextu- defined levels of partial similarity. While alizing text. While STS has previously just involved prior evaluations constrained themselves to judging text snippets that are written in the same lan- just monolingual snippets of text, the 2016 guage, this year’s evaluation includes a pilot subtask shared task includes a pilot subtask on com- on the evaluation of cross-lingual sentence pairs. puting semantic similarity on cross-lingual text snippets. This year’s traditional mono- The systems and techniques explored as a part of lingual subtask involves the evaluation of En- STS have a broad range of applications including glish text snippets from the following four do- Machine Translation (MT), Summarization, Gener- mains: Plagiarism Detection, Post-Edited Ma- ation and Question Answering (QA).
    [Show full text]
  • Semeval 2017 Task 10: Scienceie - Extracting Keyphrases and Relations from Scientific Publications
    SemEval 2017 Task 10: ScienceIE - Extracting Keyphrases and Relations from Scientific Publications Isabelle Augenstein1, Mrinal Das2, Sebastian Riedel1, Lakshmi Vikraman2, and Andrew McCallum2 1Department of Computer Science, University College London (UCL), UK 2College of Information and Computer Sciences, University of Massachusetts Amherst, USA Abstract classification and relation extraction. However, keyphrases are much more challenging to identify We describe the SemEval task of extract- than e.g. person names, since they vary signifi- ing keyphrases and relations between them cantly between domains, lack clear signifiers and from scientific documents, which is cru- contexts and can consist of many tokens. For this cial for understanding which publications purpose, a double-annotated corpus of 500 pub- describe which processes, tasks and ma- lications with mention-level annotations was pro- terials. Although this was a new task, we duced, consisting of scientific articles of the Com- had a total of 26 submissions across 3 eval- puter Science, Material Sciences and Physics do- uation scenarios. We expect the task and mains. the findings reported in this paper to be Extracting keyphrases and relations between relevant for researchers working on under- them is of great interest to scientific publishers as standing scientific content, as well as the it helps to recommend articles to readers, high- broader knowledge base population and light missing citations to authors, identify poten- information extraction communities. tial reviewers for submissions, and analyse re- search trends over time. Note that organising 1 Introduction keyphrases in terms of synonym and hypernym re- Empirical research requires gaining and maintain- lations is particularly useful for search scenarios, ing an understanding of the body of work in spe- e.g.
    [Show full text]
  • A Machine Translation Approach to Cross-Lingual Word Sense Disambiguation (Semeval-2013 Task 10)
    NRC: A Machine Translation Approach to Cross-Lingual Word Sense Disambiguation (SemEval-2013 Task 10) Marine Carpuat National Research Council Ottawa, Canada [email protected] Abstract (WSD) tasks: as a tool to automatically create train- ing data (Guo and Diab, 2010, for instance) ; as This paper describes the NRC submission to a source of parallel data that can be used to train the Spanish Cross-Lingual Word Sense Dis- ambiguation task at SemEval-2013. Since this WSD systems (Ng and Chan, 2007; van Gompel, word sense disambiguation task uses Spanish 2010; Lefever et al., 2011); or as an application translations of English words as gold annota- which can use the predictions of WSD systems de- tion, it can be cast as a machine translation veloped for SemEval tasks (Carpuat and Wu, 2005; problem. We therefore submitted the output of Chan et al., 2007; Carpuat and Wu, 2007). This Se- a standard phrase-based system as a baseline, mEval shared task gives us the opportunity to com- and investigated ways to improve its sense dis- pare the performance of machine translation systems ambiguation performance. Using only local context information and no linguistic analy- with other submissions which use very different ap- sis beyond lemmatization, our machine trans- proaches. Our goal is to provide machine transla- lation system surprisingly yields top precision tion output which is representative of state-of-the-art score based on the best predictions. However, approaches, and provide a basis for comparing its its top 5 predictions are weaker than those strength and weaknesses with that of other systems from other systems.
    [Show full text]
  • L2 Translation Assistance by Emulating the Manual Post-Editing Process
    Sensible: L2 Translation Assistance by Emulating the Manual Post-Editing Process Liling Tan, Anne-Kathrin Schumann, Jose M.M. Martinez and Francis Bond1 Universitat¨ des Saarland / Campus, Saarbrucken,¨ Germany Nanyang Technological University1 / 14 Nanyang Drive, Singapore [email protected], [email protected], [email protected], [email protected] Abstract The main difference is that in the context of post- editing the source text is provided. A transla- This paper describes the Post-Editor Z sys- tion workflow that incorporates post-editing be- tem submitted to the L2 writing assis- gins with a source sentence, e.g. “I could still tant task in SemEval-2014. The aim of sit on the border in the very last tram to West task is to build a translation assistance Berlin.” and the human translator is provided with system to translate untranslated sentence a machine-generated translation with untranslated fragments. This is not unlike the task fragments such as the previous example and some- of post-editing where human translators times “fixing” the translation would simply re- improve machine-generated translations. quire substituting the appropriate translation for Post-Editor Z emulates the manual pro- the untranslated fragment. cess of post-editing by (i) crawling and ex- tracting parallel sentences that contain the 2 Related Tasks and Previous untranslated fragments from a Web-based Approaches translation memory, (ii) extracting the pos- sible translations of the fragments indexed The L2 writing assistant task lies between the by the translation memory and (iii) apply- lines of machine translation and crosslingual word ing simple cosine-based sentence similar- sense disambiguation (CLWSD) or crosslingual ity to rank possible translations for the un- lexical substitution (CLS) (Lefever and Hoste, translated fragment.
    [Show full text]
  • SAARSHEFF at Semeval-2016 Task 1: Semantic Textual Similarity With
    SAARSHEFF at SemEval-2016 Task 1: Semantic Textual Similarity with Machine Translation Evaluation Metrics and (eXtreme) Boosted Tree Ensembles Liling Tan1, Carolina Scarton2, Lucia Specia2 and Josef van Genabith1, 3 Universitat¨ des Saarlandes1 / Campus A2.2, Saarbrucken,¨ Germany University of Sheffield2 / Regent Court, 211 Portobello, Sheffield, UK Deutsches Forschungszentrum fur¨ Kunstliche¨ Intelligenz3 / Saarbrucken,¨ Germany [email protected], [email protected], [email protected], josef.van [email protected] Abstract ing results (Agirre et al., 2012; Agirre et al., 2013; Agirre et al., 2014; Agirre et al., 2015). This paper describes the SAARSHEFF sys- At the pilot English STS-2012 task, Rios et al. tems that participated in the English Seman- (2012) trained a Support Vector Regressor using the tic Textual Similarity (STS) task in SemEval- lexical overlaps between the surface strings, named 2016. We extend the work on using ma- entities and semantic role labels and the BLEU (Pa- chine translation (MT) metrics in the STS task by automatically annotating the STS datasets pineni et al., 2002) and METEOR (Banerjee and with a variety of MT scores for each pair of Lavie, 2005; Denkowski and Lavie, 2010) scores be- text snippets in the STS datasets. We trained tween the text snippets and their best system scored our systems using boosted tree ensembles and a Pearson correlation mean of 0.3825. The system achieved competitive results that outperforms underperformed compared to the organizers’ base- he median Pearson correlation scores from all line system1 which scored 0.4356. participating systems. For the English STS-2013 task, Barron-Cede´ no˜ et al.
    [Show full text]